The writing, reading and chromatin signatures of histone H3K23 on mammalian chromatin

The DNA of eukaryotes is wrapped around histone proteins to form the nucleosome, the fundamental repeating unit of chromatin. Post-translational modifications (PTMs) on histone proteins can act to regulate chromatin compaction by shifting towards a permissive/open state in which DNA is accessible to the environment (euchromatin) or towards a transcriptionally repressive/closed state that restricts access of the DNA to the environment (heterochromatin). Lysine methylation is a prominent histone PTM that has been implicated in the formation of both repressive and permissive chromatin states. Methylation of histone H3K23, a novel yet poorly understood class of histone methylation, has proven critical for preserving genomic integrity in T. thermophila, maintaining RNAi pathways in C. elegans, and transposon silencing in A. thaliana, but very little information is known about the function of H3K23 methylation on mammalian chromatin. Here, we performed in vitro histone methyltransferase assays to screen canonical H3K9 writers, and identified EHMT1/GLP and EHMT2/G9a as writers of H3K23 methylation. To follow up, we used a combination of in vitro and in vivo tools to perturb both enzymes at the protein and DNA level, and we identified differential activity between the two enzymes on H3K23. While both enzymes can mediate mono- and di-methylation of histone H3K23, only EHMT1/GLP can catalyze tri-methylation of H3K23. Interestingly, our studies also identified H3K18 as a target for mono-, di- and tri-methylation, de novo, by both EHMT1/GLP and EHMT2/G9a in vitro and in vivo. Additionally, we performed nChIP-Seq on H3K23me3 and found it to be a promoter mark that is often heavily bivalent with H3K4me3 at promoters. Our data also shows that monovalent H3K23me3 occupies transcription start sites (TSSs) while bivalent H3K23me3 (with H3K4me3) occupies gene promoters. Additionally, we have evidence that H3K23me3 dampens gene expression relative to H3K4me3, suggesting that H3K23me3 might serve to silence gene expression on mammalian chromatin. Lastly, we found that perturbations in H3K4me3 effect H3K23me3, but not vice versa. This suggests unique one directional cross-talk between H3K4me3 and H3K23me3 and future studies should aim to elucidate the functional interplay of these bivalent chromatin modifications